# -*- coding: utf-8 -*-
# Copyright (c) 2015, Alex Grigorevskiy, Arno Solin
# Licensed under the BSD 3-clause license (see LICENSE.txt)
"""
Classes in this module enhance several stationary covariance functions with the
Stochastic Differential Equation (SDE) functionality.
"""
from .rbf import RBF
from .stationary import Exponential
from .stationary import RatQuad
import numpy as np
import scipy as sp
try:
from scipy.linalg import solve_continuous_lyapunov as lyap
except ImportError:
from scipy.linalg import solve_lyapunov as lyap
import warnings
[docs]class sde_RBF(RBF):
"""
Class provide extra functionality to transfer this covariance function into
SDE form.
Radial Basis Function kernel:
.. math::
k(r) = \sigma^2 \exp \\bigg(- \\frac{1}{2} r^2 \\bigg) \\ \\ \\ \\ \text{ where } r = \sqrt{\sum_{i=1}^{input dim} \frac{(x_i-y_i)^2}{\ell_i^2} }
"""
def __init__(self, *args, **kwargs):
"""
Init constructior.
Two optinal extra parameters are added in addition to the ones in
RBF kernel.
:param approx_order: approximation order for the RBF covariance. (Default 10)
:type approx_order: int
:param balance: Whether to balance this kernel separately. (Defaulf True). Model has a separate parameter for balancing.
:type balance: bool
"""
if 'balance' in kwargs:
self.balance = bool( kwargs.get('balance') )
del kwargs['balance']
else:
self.balance = True
if 'approx_order' in kwargs:
self.approx_order = kwargs.get('approx_order')
del kwargs['approx_order']
else:
self.approx_order = 6
super(sde_RBF, self).__init__(*args, **kwargs)
[docs] def sde_update_gradient_full(self, gradients):
"""
Update gradient in the order in which parameters are represented in the
kernel
"""
self.variance.gradient = gradients[0]
self.lengthscale.gradient = gradients[1]
[docs] def sde(self):
"""
Return the state space representation of the covariance.
Note! For Sparse GP inference too small or two high values of lengthscale
lead to instabilities. This is because Qc are too high or too low
and P_inf are not full rank. This effect depends on approximatio order.
For N = 10. lengthscale must be in (0.8,8). For other N tests must be conducted.
N=6: (0.06,31)
Variance should be within reasonable bounds as well, but its dependence is linear.
The above facts do not take into accout regularization.
"""
#import pdb; pdb.set_trace()
if self.approx_order is not None:
N = self.approx_order
else:
N = 10# approximation order ( number of terms in exponent series expansion)
roots_rounding_decimals = 6
fn = np.math.factorial(N)
p_lengthscale = float( self.lengthscale )
p_variance = float(self.variance)
kappa = 1.0/2.0/p_lengthscale**2
Qc = np.array( ((p_variance*np.sqrt(np.pi/kappa)*fn*(4*kappa)**N,),) )
eps = 1e-12
if (float(Qc) > 1.0/eps) or (float(Qc) < eps):
warnings.warn("""sde_RBF kernel: the noise variance Qc is either very large or very small.
It influece conditioning of P_inf: {0:e}""".format(float(Qc)) )
pp1 = np.zeros((2*N+1,)) # array of polynomial coefficients from higher power to lower
for n in range(0, N+1): # (2N+1) - number of polynomial coefficients
pp1[2*(N-n)] = fn*(4.0*kappa)**(N-n)/np.math.factorial(n)*(-1)**n
pp = sp.poly1d(pp1)
roots = sp.roots(pp)
neg_real_part_roots = roots[np.round(np.real(roots) ,roots_rounding_decimals) < 0]
aa = sp.poly1d(neg_real_part_roots, r=True).coeffs
F = np.diag(np.ones((N-1,)),1)
F[-1,:] = -aa[-1:0:-1]
L= np.zeros((N,1))
L[N-1,0] = 1
H = np.zeros((1,N))
H[0,0] = 1
# Infinite covariance:
Pinf = lyap(F, -np.dot(L,np.dot( Qc[0,0],L.T)))
Pinf = 0.5*(Pinf + Pinf.T)
# Allocating space for derivatives
dF = np.empty([F.shape[0],F.shape[1],2])
dQc = np.empty([Qc.shape[0],Qc.shape[1],2])
dPinf = np.empty([Pinf.shape[0],Pinf.shape[1],2])
# Derivatives:
dFvariance = np.zeros(F.shape)
dFlengthscale = np.zeros(F.shape)
dFlengthscale[-1,:] = -aa[-1:0:-1]/p_lengthscale * np.arange(-N,0,1)
dQcvariance = Qc/p_variance
dQclengthscale = np.array(( (p_variance*np.sqrt(2*np.pi)*fn*2**N*p_lengthscale**(-2*N)*(1-2*N),),))
dPinf_variance = Pinf/p_variance
lp = Pinf.shape[0]
coeff = np.arange(1,lp+1).reshape(lp,1) + np.arange(1,lp+1).reshape(1,lp) - 2
coeff[np.mod(coeff,2) != 0] = 0
dPinf_lengthscale = -1/p_lengthscale*Pinf*coeff
dF[:,:,0] = dFvariance
dF[:,:,1] = dFlengthscale
dQc[:,:,0] = dQcvariance
dQc[:,:,1] = dQclengthscale
dPinf[:,:,0] = dPinf_variance
dPinf[:,:,1] = dPinf_lengthscale
P0 = Pinf.copy()
dP0 = dPinf.copy()
if self.balance:
# Benefits of this are not very sound. Helps only in one case:
# SVD Kalman + RBF kernel
import GPy.models.state_space_main as ssm
(F, L, Qc, H, Pinf, P0, dF, dQc, dPinf,dP0) = ssm.balance_ss_model(F, L, Qc, H, Pinf, P0, dF, dQc, dPinf, dP0 )
return (F, L, Qc, H, Pinf, P0, dF, dQc, dPinf, dP0)
[docs]class sde_Exponential(Exponential):
"""
Class provide extra functionality to transfer this covariance function into
SDE form.
Exponential kernel:
.. math::
k(r) = \sigma^2 \exp \\bigg(- \\frac{1}{2} r \\bigg) \\ \\ \\ \\ \text{ where } r = \sqrt{\sum_{i=1}^{input dim} \frac{(x_i-y_i)^2}{\ell_i^2} }
"""
[docs] def sde_update_gradient_full(self, gradients):
"""
Update gradient in the order in which parameters are represented in the
kernel
"""
self.variance.gradient = gradients[0]
self.lengthscale.gradient = gradients[1]
[docs] def sde(self):
"""
Return the state space representation of the covariance.
"""
variance = float(self.variance.values)
lengthscale = float(self.lengthscale)
F = np.array(((-1.0/lengthscale,),))
L = np.array(((1.0,),))
Qc = np.array( ((2.0*variance/lengthscale,),) )
H = np.array(((1.0,),))
Pinf = np.array(((variance,),))
P0 = Pinf.copy()
dF = np.zeros((1,1,2));
dQc = np.zeros((1,1,2));
dPinf = np.zeros((1,1,2));
dF[:,:,0] = 0.0
dF[:,:,1] = 1.0/lengthscale**2
dQc[:,:,0] = 2.0/lengthscale
dQc[:,:,1] = -2.0*variance/lengthscale**2
dPinf[:,:,0] = 1.0
dPinf[:,:,1] = 0.0
dP0 = dPinf.copy()
return (F, L, Qc, H, Pinf, P0, dF, dQc, dPinf, dP0)
[docs]class sde_RatQuad(RatQuad):
"""
Class provide extra functionality to transfer this covariance function into
SDE form.
Rational Quadratic kernel:
.. math::
k(r) = \sigma^2 \\bigg( 1 + \\frac{r^2}{2} \\bigg)^{- \alpha} \\ \\ \\ \\ \text{ where } r = \sqrt{\sum_{i=1}^{input dim} \frac{(x_i-y_i)^2}{\ell_i^2} }
"""
[docs] def sde(self):
"""
Return the state space representation of the covariance.
"""
assert False, 'Not Implemented'
# Params to use:
# self.lengthscale
# self.variance
#self.power
#return (F, L, Qc, H, Pinf, dF, dQc, dPinf)